CN113037162B

CN113037162B - Vibration compensation controller for neural network band-pass filter of bearingless permanent magnet synchronous motor

Info

Publication number: CN113037162B
Application number: CN202110195977.3A
Authority: CN
Inventors: 朱熀秋; 王鑫
Original assignee: Jiangsu University
Current assignee: Jiangsu University
Priority date: 2021-02-22
Filing date: 2021-02-22
Publication date: 2022-04-26
Anticipated expiration: 2041-02-22
Also published as: CN113037162A; US20230008153A1; US11705838B2; WO2022174488A1

Abstract

The invention discloses a vibration compensation controller of a neural network band-pass filter of a bearingless permanent magnet synchronous motor, which consists of a displacement controller and a rotating speed controller, wherein the displacement controller comprises a vibration force compensation control module and a dead zone vibration compensation module; the vibration force compensation control module takes actual displacement and a rotor mechanical angle as input, outputs corresponding vibration compensation force and consists of a first neural network band-pass filter, a second neural network band-pass filter, a third PID controller and a fourth PID controller; the dead zone vibration compensation module takes the rotor electrical angle and the actual alternating current and direct current as input, outputs alternating-direct axis compensation voltage, and consists of a third neural network band-pass filter in the direction of a direct axis, a fourth neural network band-pass filter in the direction of an alternating axis, a sixth PI controller and a seventh PI controller; the invention not only analyzes and compensates the vibration caused by the eccentricity problem, but also compensates and controls the vibration caused by the dead zone effect, thereby effectively improving the suspension control precision.

Description

Vibration compensation controller for neural network band-pass filter of bearingless permanent magnet synchronous motor

Technical Field

The invention belongs to the field of control of a bearingless permanent magnet synchronous motor, and relates to a technology for dead zone compensation control and rotor eccentricity control of the bearingless permanent magnet synchronous motor, which is used for performing compensation control on vibration of the bearingless permanent magnet synchronous motor.

Background

The bearingless permanent magnet synchronous motor is a novel special motor with high rotating speed, high precision and no need of lubrication, and has wider and wider application prospect in aerospace aviation, chemical manufacturing, semiconductor industry and other fields requiring special environments. The bearingless permanent magnet synchronous motor is used as a rotary driving motor, and the problems of uneven material, processing error, assembly error and the like inevitably cause rotor mass eccentricity to a certain degree and generate centrifugal exciting force with the same frequency of rotating speed during rotation. Meanwhile, in the control process of the bearingless permanent magnet synchronous motor, dead time is required to be set to avoid short circuit of upper and lower bridge arms of the inverter, and the introduction of the dead time increases current harmonics, further increases the amplitude of unbalanced force, causes unbalanced vibration of the rotor, and affects the suspension control precision of the rotor.

Regarding the control of the unbalanced vibration of the rotor of the bearingless permanent magnet synchronous motor, the prior art mostly performs compensation control on the unbalanced vibration caused by the eccentricity of the rotor mass, and the unbalanced vibration caused by the dead zone effect is mentioned. Chinese patent publication No. CN104659990A discloses a self-adaptive filter unbalanced vibration displacement extraction method for a bearingless motor, which lays a foundation for the primary condition of vibration compensation control of the bearingless motor. Chinese patent publication No. CN105048913A discloses a current compensation-based unbalanced vibration control system for an asynchronous motor without bearing, which realizes suspension vibration compensation control by adjusting compensation current. However, in these schemes, the vibration compensation control of the bearingless motor is mainly performed by detecting and compensating the vibration caused by eccentricity, but the vibration caused by the dead zone effect is not mentioned yet. In order to improve the control precision of the unbalanced vibration displacement of the bearingless permanent magnet synchronous motor, the compensation is required to be carried out on the rotor eccentric displacement caused by the rotor mass eccentricity and the rotor unbalanced vibration caused by the dead zone effect, which is the important factor for realizing the control of the high-precision bearingless permanent magnet synchronous motor.

Disclosure of Invention

The invention aims to provide a vibration compensation controller of a neural network band-pass filter of a bearingless permanent magnet synchronous motor, which can inhibit vibration compensation of vibration of the bearingless permanent magnet synchronous motor, and solves the problem that the existing bearingless permanent magnet synchronous motor only carries out vibration compensation on rotor mass eccentricity in vibration compensation control and ignores vibration caused by a dead zone effect, so that stable suspension and efficient operation of a motor rotor are realized, the motor control precision is improved, and the vibration compensation controller is better applied to an electric transmission system.

The invention provides a vibration compensation controller of a bearingless permanent magnet synchronous motor neural network band-pass filter, which adopts the technical scheme that: the device comprises a displacement controller and a rotating speed controller, wherein the displacement controller comprises a vibration force compensation control module and a dead zone vibration compensation module;

the vibration force compensation control module uses actual displacement x, y in x, y directions and a rotor mechanical angle theta_mAs input, a corresponding vibration compensation force F is output_xh，F_yhThe system consists of a first neural network band-pass filter, a second neural network band-pass filter, a third PID controller and a fourth PID controller; the first neural network band-pass filter is used for measuring the actual displacement x and the mechanical angle theta of the rotor in the x direction_mAs input, output vibrational displacement

Given value of 0 and vibration displacement

Taking the difference as the input of a third PID controller, which outputs a vibration compensation force F_xh(ii) a The actual displacement y of the second neural network band-pass filter in the y direction and the mechanical angle theta of the rotor_mAs input, output vibrational displacement

Given value of 0 and vibration displacement

Taking the difference as the input of the fourth PID controller, the fourth PID controller outputs the vibration compensation force F_yh(ii) a Said vibration compensation force F_xhGiven value F of force in x direction of levitation winding_xAfter summing, inputting the sum to a force current conversion module; said vibration compensation force F_yhGiven value F of force in y direction of levitation winding_yThe summed signals are input to a force current conversion module, and the current conversion module obtains a given value of the quadrature-direct axis current

And

the dead zone vibration compensation module uses a rotor electrical angle theta_eAnd the actual current i of the quadrature-direct axis current_Bq，i_BdAs input, the AC-DC axis compensation voltage u is output_Bqh，u_BdhThe device consists of a third neural network band-pass filter in the direction of a direct axis, a fourth neural network band-pass filter in the direction of a quadrature axis, a sixth PI controller and a seventh PI controller, wherein the third neural network band-pass filter uses actual current i in the direction of the direct axis_BdElectrical angle theta with rotor_e6 times of the current is used as input to obtain harmonic current in the direction of a straight shaft

With 0 as given value and harmonic current

Taking the difference and using the result as the input of a sixth PI controller, which obtains a direct axis compensation voltage u_BdhThe control voltage u in the direction of the straight axis_BdWith compensation voltage u of the direct axis_BdhAdding to obtain direct axis command voltage

The fourth neural network band-pass filter is used for measuring the actual current i in the quadrature axis direction_BqElectrical angle theta with rotor_e6 times of the current is used as input to obtain harmonic current in the direction of a straight shaft

With 0 as given value and harmonic current

Taking the difference and the result as the input of a seventh PI controller, which obtains the quadrature axis compensation voltage u_BqhWill control the voltage u in the quadrature axis direction_BqAnd quadrature axis compensation voltage u_BqhAdding to obtain quadrature axis command voltage

The invention has the beneficial effects that:

1) the dead zone vibration compensation control is adopted, so that the dead zone is compensated, the vibration of the bearingless permanent magnet synchronous motor in the running process can be effectively inhibited, and the suspension control precision is improved.

2) The neural network band-pass filter adopted by the invention has simple working principle and concise calculation process, and can acquire the required signal according to the real-time rotating speed of the motor.

3) The invention adopts the PI controller to adjust the vibration, and the controller has simple principle, convenient coefficient adjustment and stronger robustness.

4) In the vibration compensation control of the bearingless permanent magnet synchronous motor, the compensation control is generally carried out by only considering the vibration caused by the eccentricity factor, but the vibration problem caused by the dead zone effect is not mentioned, which is disadvantageous to the accuracy of the whole levitation control. In order to enable the bearingless permanent magnet synchronous motor to have higher suspension control precision, the vibration caused by the eccentricity problem is analyzed and compensated, the vibration caused by the dead zone effect is compensated and controlled, and the suspension control precision is effectively improved.

Drawings

In order that the invention may be more fully understood, reference is now made to the following detailed description of the invention taken in conjunction with the accompanying drawings in which:

FIG. 1 is a schematic block diagram of the architecture of the present invention;

FIG. 2 is a schematic block diagram of the structure of the rotational speed controller 2 in FIG. 1;

FIG. 3 is a schematic block diagram of the structure of the displacement controller 1 in FIG. 1;

fig. 4 is a schematic block diagram of the vibration force compensation module 5 in the x-direction and the y-direction in fig. 3;

FIG. 5 is a schematic block diagram of the dead zone vibration compensation module 6 in the direct axis direction and the quadrature axis direction of FIG. 3;

fig. 6 is a functional block diagram of the internal structure of the first neural network band-pass filter 51 in fig. 4;

FIG. 7 is a schematic block diagram of the internal structure of the second neural network band-pass filter 53 in FIG. 5;

fig. 8 is a functional block diagram of the internal structure of the third neural network band-pass filter 61 in fig. 6;

fig. 9 is a functional block diagram of the internal structure of the fourth neural network band-pass filter 63 in fig. 7;

fig. 10 is a schematic block diagram of a general implementation of the structure of the vibration compensation controller of the motor according to the present invention.

In the figure: 1. a displacement controller; 2. a rotational speed controller; 3. a bearingless permanent magnet synchronous motor; 11. a first PID controller; 12. a second PID controller; 13. a force-current conversion module; 14. a fourth PI controller; 15. a fifth PI controller; 16. a third coordinate transformation module; 17. an angle calculation module; 21. a first PI controller; 22. a second PI controller; 23. a third PI controller; 24. a first coordinate transformation module; 25. a second coordinate transformation module; 26. a first SVPWM inverter; 27. an encoder; 28. a speed calculation module; 5. a vibration force compensation module; 51. a first neural network band pass filter; 52. a third PID controller; 53. a second neural network band pass filter; 54. a fourth PID controller; 55. a first weight value adjusting module; 56. a second weight value adjusting module; 6. a dead zone vibration compensation module; 61. a third neural network band-pass filter; 62. a sixth PI controller; 63. a fourth neural network band-pass filter; 64. a seventh PI controller; 65. a third weight value adjusting module; 66. a fourth weight value adjusting module; 90. a second SVPWM inverter; 91. a fourth coordinate transformation module; 92. and a displacement calculation module.

Detailed Description

The specific idea and implementation steps of the invention are as follows:

referring to fig. 1, the vibration compensation controller of the neural network band-pass filter of the bearingless permanent magnet synchronous motor is composed of a displacement controller 1 and a rotating speed controller 2, and the output ends of the displacement controller 1 and the rotating speed controller 2 are connected with a bearingless permanent magnet synchronous motor 3 to control the bearingless permanent magnet synchronous motor 3.

As for the rotation speed controller 2, as shown in fig. 2, it employs a speed current double closed loop control, which is composed of a first PI controller 21, a second PI controller 22, a third PI controller 23, a first coordinate transformation module 24, a second coordinate transformation module 25, a first SVPWM inverter 26, an encoder 27, and a speed calculation module 28. The output end of the encoder 27 is connected to the speed calculation module 28, the encoder 27 collects the rotating speed pulse signals from the rotating shaft of the bearingless permanent magnet synchronous motor 3 and performs accumulation operation, the accumulated result Δ P is input to the speed calculation module 28, the actual rotating speed n of the motor rotor is obtained through calculation of the speed calculation module 28, and the calculating formula of the rotating speed n is as follows:

in the formula: t is_sAn interrupt period of the rotational speed controller 2; l is_eIs the number of lines of the encoder.

The calculated actual rotating speed n and the rotating speed given value n^*Making difference to obtain rotation speed error, inputting the error into the first PI controller 21, and obtaining the given value of the quadrature axis current of the torque winding after being regulated by the first PI controller 21

Meanwhile, a current sensor is used for collecting the torque current i of the torque two-phase winding of the bearingless permanent magnet synchronous motor 3_2AAnd i_2CApplying a torque current i_2AAnd i_2CInputting the data into a second coordinate transformation module 25, wherein the second coordinate transformation module 25 consists of Clarke transformation and Park transformation, and i is subjected to the second coordinate transformation module 25_2AAnd i_2CObtaining the actual value i of the quadrature axis current of the torque winding under the rotating coordinate system after conversion_MqAnd the actual value i of the direct axis current of the torque winding_Md. Given value of quadrature-axis current of torque winding

Actual value i of quadrature axis current of sum torque winding_MqIs used as the input of the second PI controller 22 to obtain the given value of the quadrature axis voltage of the torque winding

By torque winding of direct axis current

As a given value, will

And the actual value i of the direct axis current of the torque winding_MdIs used as the input of the third PI controller 23 to obtain the given value of the direct-axis voltage of the torque winding

The output ends of the second PI controller 22 and the third PI controller 23 are connected with the input end of a first coordinate transformation module 24, the first coordinate transformation module 24 is composed of Park inverse transformation, and the transformation can enable the quadrature axis voltage given value of the torque winding to be converted into the quadrature axis voltage given value

And torque winding direct axis voltage set point

Converting into torque winding voltage u under static coordinate system_MαAnd u_Mβ. The output end of the first coordinate transformation module 24 is sequentially connected with the first SVPWM inverter 26 and the bearingless permanent magnet synchronous motor 3 in series, and the first coordinate transformation module 24 transforms the voltage u_MαAnd u_MβThe output of the first SVPWM inverter 26 is connected with the input of the bearingless permanent magnet synchronous motor 3 as the input of the first SVPWM inverter 26, and the three-phase input voltage u of the bearingless permanent magnet synchronous motor 3 is obtained through the first SVPWM inverter 26_2A、u_2B、u_2C。

As for the displacement controller 1, as shown in fig. 3, it employs displacement current double closed loop control, which is composed of a first PID controller 11, a second PID controller 12, a vibration force compensation module 5, a force current conversion module 13, a fourth PI controller 14, a fifth PI controller 15, a dead zone vibration compensation module 6, a third coordinate conversion module 16, an angle calculation module 17, a second SVPWM inverter 90, a fourth coordinate conversion module 91, a displacement calculation module 92, and an encoder 27. Wherein, the rotor position of the bearingless permanent magnet synchronous motor 3 is collected by a displacement sensor and is input into displacement calculationA module 92, the displacement calculating module 92 converts the collected displacement signals into actual displacements in x and y directions, and the actual displacement x in the x direction and a given value x are calculated^*Making difference to obtain displacement error, inputting the displacement error into the first PID controller 11, and obtaining the given value F of the force in the x direction of the suspension winding after being regulated by the first PID controller 11_x(ii) a The actual displacement y in the y direction is compared with a given value y^*The difference is made to obtain a displacement error, the displacement error is input into the second PID controller 82, and the given value F of the force in the y direction of the suspension winding is obtained after the adjustment of the second PID controller 12_y。

The output end of the encoder 27 is further connected with an angle calculation module 17, and the pulse signal output by the encoder 27 passes through the angle calculation module 17 to obtain the mechanical angle theta of the rotor_mAnd the calculation process of the mechanical angle of the rotor at the moment k is as follows:

in the formula: Δ P is the accumulated result of the pulses output from encoder 27.

The output ends of the angle calculation module 17 and the displacement calculation module 92 are both connected with the input end of the vibration force compensation control module 5, and the vibration force compensation control module 5 calculates the mechanical angle theta of the rotor output by the angle calculation module 17_mAnd the actual rotor displacement x, y output by the displacement calculation module 92 is used as input to obtain the compensation force F_xhAnd F_yh。

As shown in fig. 4, the vibration force compensation control module 5 is composed of a first neural network band-pass filter 51, a second neural network band-pass filter 53, a third PID controller 52, and a fourth PID controller 54. By displacement in the x-direction and rotor mechanical angle theta_mAs an input to the first neural network band pass filter 51, the output is the vibration displacement

A signal. The specific structure of the first neural network band-pass filter 51 in the x direction is shown in fig. 6, and includes a first weight adjusting module 5 for adjusting the actual displacement x and the first nerveVibration displacement of the output of the network band-pass filter 51

Making a difference to obtain an error signal e_xWill error signal e_xMechanical angle theta to rotor_mAs an input of the first weight adjustment module 55, to obtain an updated weight ω in the x direction_{x_1}And ω_{x_2}. Vibration displacement output by the first neural network band-pass filter 51

The calculation formula at time k is:

weight omega_{x_1}And ω_{x_2}The calculation process of (2) adopts the following formula:

in the formula: e.g. of the type_xFiltering the harmonic components in the x direction; omega_{x_1}，ω_{x_2}The weight value updated in the x direction; mu.s₁Is the step size factor.

Thereby obtaining the vibration displacement in the x direction

As shown in FIG. 4, the given value and the vibration displacement are 0

Making a difference, using the result of the displacement difference as the input of the third PID controller 52, and obtaining the vibration compensation force F after being adjusted by the third PID controller 52_xh。

The second neural network band-pass filter 53 has the same structure and principle as the first neural network band-pass filter 51. Similarly, the displacement in the y direction and the mechanical angle theta of the rotor_mAs an input of the second neural network band pass filter 53, the specific structure of the second neural network band pass filter 53 in the y direction is shown in FIG. 7, and the actual displacement y and the vibration displacement output from the second neural network band pass filter 53 are set

Making a difference to obtain an error signal e_yAnd apply the error signal e_yMechanical angle theta to rotor_mThe sine and cosine values of the first weight value are used as the input of the second weight value adjusting module 56, so as to obtain the updated weight value omega in the y direction_{y_1}And ω_{y_2}. The vibration displacement signal of k time output by the second neural network band-pass filter 53

The calculation formula of (2) is as follows:

weight omega_{y_1}And ω_{y_2}The calculation process of (2) adopts the following formula:

in the formula: e.g. of the type_yFiltering the harmonic component in the y direction; omega_{y_1}，ω_{y_2}The weight value updated in the y direction; mu.s₁Is the step size factor.

Thereby obtaining a vibration displacement signal in the y direction

As shown in FIG. 4, the given value and the vibration displacement signal are 0

Making a difference, using the result of the displacement difference as the input of the fourth PID controller 54, and obtaining the vibration compensation force F after being adjusted by the fourth PID controller 54_yh。

The force F in the x-direction to be output by the first PID controller 11_xThe vibration compensation force F in the x direction output by the vibration force compensation module 5_xhSummed with the force F in the y-direction output by the second PID controller 12_yThe vibration compensation force F in the y direction output by the vibration force compensation module 5_yhAfter summing, the sum is input into the power current conversion module 13, so as to obtain the alternating current and direct current given value of the suspension winding

And

setting the obtained quadrature-direct axis current

And

actual current i of quadrature-direct axis current of suspension winding_BqAnd i_BdThe differences were made separately. Wherein i_BqAnd i_BdCollecting the two-phase suspension winding current of the bearingless permanent magnet synchronous motor 3 through a current sensor, and collecting the current i_1AAnd i_1CInputting the data into a fourth coordinate transformation module 91, wherein the fourth coordinate transformation module 91 consists of Clarke transformation and Park transformation, i_1AAnd i_1CThe actual current i of the suspension winding quadrature-direct axis can be obtained through the fourth coordinate transformation module 91_BqAnd i_BdWill be

And i_BqThe difference result is inputted to the fourth PI controller 14, so as to obtain the suspension winding quadrature axis control voltage u_Bq(ii) a Will be provided with

And i_BdThe difference result is inputted to the fifth PI controller 15, and the suspension winding is obtainedShaft control voltage u_Bd。

At rotor electrical angle theta_eAnd actual value i of quadrature axis current of suspension winding_BqActual value i of direct axis current of suspension winding_BdInput into the dead zone vibration compensation module 6 to obtain a compensation voltage u_BqhAnd u_Bdh. Wherein, the pulse signal of the bearingless permanent magnet synchronous motor 3 collected by the encoder 27 passes through the angle calculation module 17 to obtain the rotor electrical angle theta_eThe calculation process is as follows:

θ_e(k)＝P_Mθ_m(k) (7)

in the formula: theta_m(k) The mechanical angle of the rotor at the moment k in the formula (2); p_MIs the torque winding pole pair number.

The obtained rotor electrical angle theta_eAnd the actual value i of the quadrature-direct axis current_BqAnd i_BdAnd the signals are input to the dead zone vibration compensation module 6, and the dead zone vibration compensation module 6 is composed of a third neural network band-pass filter 61 in the direction of a direct axis, a fourth neural network band-pass filter 63 in the direction of a quadrature axis, a sixth PI controller 62 and a seventh PI controller 64. In the dead zone vibration compensation module 6, compensation in the direct axis direction and the quadrature axis direction is as shown in fig. 5. Current i in the direction of the straight axis_BdElectrical angle theta with rotor_eThe 6 times of the harmonic current is used as the input of the third neural network band-pass filter 61 in the direction of the direct axis, so that the harmonic current signal in the direction of the direct axis is obtained

Fig. 8 shows a schematic diagram of an internal structure of the third neural network band-pass filter 61 in the direction of the straight axis, which includes a third weight value adjusting module 65. In FIG. 8, the current i in the direction of the straight axis is measured_BdAnd the harmonic current signal output by the third neural network band-pass filter 61

Difference is made to obtain an error signal e_BdWill error signal e_BdElectrical angle theta with rotor_eThe sine and cosine value of 6 times is used as the input of the third weight value adjusting module 65, so as to obtain the updated straight axis directionWeight value omega of_{d6_1}And ω_{d6_2}. Harmonic current at time k output from the third neural network band-pass filter 61

The calculation formula of (2) is as follows:

weight omega_{d6_1}And ω_{d6_2}The calculation process of (2) adopts the following formula:

in the formula: e.g. of the type_BdFiltering the harmonic components in the direction of the straight axis; omega_{d6_1}，ω_{d6_2}The weight of the 6 th harmonic wave updated in the direction of the straight axis; mu.s₂Is the step size factor.

Thereby obtaining harmonic current signals in the direction of the straight axis

As shown in FIG. 5, the harmonic current signal and the given value are 0

Taking the difference, and taking the result as the input of the sixth PI controller 62, and obtaining the direct axis compensation voltage u after being regulated by the sixth PI controller 62_Bdh。

The quadrature axis fourth neural network band-pass filter 63 and the third neural network band-pass filter 61 have the same structure. Similarly, with the current i in the direction of the quadrature axis_BqElectrical angle theta with rotor_eThe harmonic current signal in the direction of the direct axis is obtained by taking 6 times of the harmonic current signal as the input of the quadrature axis fourth neural network band-pass filter 63

Wherein, the internal structure schematic diagram of the quadrature direction fourth neural network band-pass filter 63As shown in fig. 9, it includes a fourth weight value adjusting module 66. In FIG. 9, the current i in the quadrature direction is measured_BqAnd the harmonic current signal output by the fourth neural network band-pass filter 63

Making a difference to obtain a current error e_BqSignal, current error e_BqSignal and rotor electrical angle theta_eThe sine and cosine value of 6 times is used as the input of the fourth weight value adjusting module 66, so as to obtain the updated weight value omega in the direction of the straight axis_{q6_1}And ω_{q6_2}. The fourth neural network band-pass filter 63 outputs harmonic current

Harmonic current at time k

The calculation formula of (2) is as follows:

in the formula: e.g. of the type_BqFiltering the harmonic components in the cross-axis direction; omega_{q6_1}And ω_{q6_2}The weight of the 6 th harmonic wave updated in the quadrature axis direction; mu.s₂Is the step size factor.

As shown in FIG. 5, the harmonic current signal and the given value are 0

Taking the difference, and using the result as the input of the seventh PI controller 64, and obtaining the quadrature axis compensation voltage u after being adjusted by the seventh PI controller 64_Bqh。

A voltage u in the direction of the straight axis to be output from the fourth PI controller 14_BdAnd a direct axis compensation voltage u output in the dead zone vibration compensation module_BdhAdding to obtain direct axis command voltage

A voltage u in the quadrature axis direction output from the fifth PI controller 15_BqQuadrature axis compensation voltage u output by dead zone vibration compensation module_BqhAdding to obtain quadrature axis command voltage

Will obtain

And

the third coordinate transformation module 16 is used as an input of the third coordinate transformation module 16, the third coordinate transformation module 16 is composed of Park inverse transformation, and the suspension winding voltage u under the static coordinate system is obtained through the third coordinate transformation module 16_BαAnd u_Bβ。

Voltage u of levitation winding_BαAnd u_BβThe output of the second SVPWM inverter 90 is connected with the input of the bearingless permanent magnet synchronous motor 3 as the input of the second SVPWM inverter 90, and the three-phase input voltage u of the bearingless permanent magnet synchronous motor 3 is obtained through the second SVPWM inverter 90_1A、u_1B、u_1C。

Fig. 10 is a general implementation schematic block diagram of the structure of the motor vibration compensation controller of the present invention, which implements speed closed-loop control and vibration compensation control by adjusting the parameters of the speed closed-loop and position closed-loop regulators through the design of the displacement controller 1, the rotational speed controller 2 and each module therein. The speed controller 2 performs speed regulation control by adopting a common vector control method with a direct axis instruction current of 0, the displacement controller 1 completes vector control by regulating displacement, so that a rotor of the bearingless permanent magnet synchronous motor 3 keeps stable operation, the vibration force compensation module 5 performs compensation control on an eccentric vibration signal in the displacement signal, and the dead zone vibration compensation control module 6 is used for performing compensation again on a higher harmonic signal in the current caused by the dead zone effect, so that more accurate vibration compensation control can be realized.

The present invention can be realized in light of the above. Other variations and modifications which may occur to those skilled in the art without departing from the spirit and scope of the invention are intended to be included within the scope of the invention.

Claims

1. The utility model provides a no bearing PMSM neural network band-pass filter vibration compensation controller which comprises displacement controller (1) and rotational speed controller (2), characterized by: the displacement controller (1) comprises a vibration force compensation control module (5) and a dead zone vibration compensation module (6);

the vibration force compensation control module (5) performs actual displacement x, y in x and y directions and a rotor mechanical angle theta_mAs input, a corresponding vibration compensation force F is output_xh，F_yhThe neural network band-pass filter comprises a first neural network band-pass filter (51), a second neural network band-pass filter (53), a third PID controller (52) and a fourth PID controller (54); the first neural network band-pass filter (51) is used for measuring the actual displacement x and the mechanical angle theta of the rotor in the x direction_mAs input, output vibrational displacement

Given value of 0 and vibration displacement

Taking the difference and using the difference as an input to a third PID controller (52), the third PID controller (52) outputting a vibration compensation force F_xh(ii) a The actual displacement y of the second neural network band-pass filter (53) in the y direction and the mechanical angle theta of the rotor_mAs input, output vibrational displacement

Given value of 0 and vibration displacement

Taking the difference as an input to a fourth PID controller (54), the fourth PID controller (54) outputting a vibration compensation force F_yh(ii) a Said vibration compensation force F_xhGiven value F of force in x direction of levitation winding_xAfter summing, the sum is input to a force current conversion module (13); said vibration compensation force F_yhGiven value F of force in y direction of levitation winding_yThe summed values are input to a force current conversion module (13), and the current conversion module (13) obtains the given value of the quadrature-direct axis current

And

the dead zone vibration compensation module (6) uses a rotor electrical angle theta_eAnd the actual current i of the quadrature-direct axis current_Bq，i_BdAs input, the AC-DC axis compensation voltage u is output_Bqh，u_BdhThe device comprises a third neural network band-pass filter (61) in the direction of a direct axis, a fourth neural network band-pass filter (63) in the direction of a quadrature axis, a sixth PI controller (62) and a seventh PI controller (64), wherein the third neural network band-pass filter (61) uses actual current i in the direction of the direct axis_BdAnd 6 times the rotor electrical angle theta_eAs input, obtaining harmonic current in the direction of the direct axis

With 0 as given value and harmonic current

Taking the difference and the result as the input of the sixth PI controller (62), the sixth PI controller (62) obtains the direct axis compensation voltageu_BdhThe control voltage u in the direction of the straight axis_BdWith compensation voltage u of the direct axis_BdhAdding to obtain direct axis command voltage

The fourth neural network band-pass filter (63) uses the actual current i in the quadrature axis direction_BqAnd 6 times the rotor electrical angle theta_eAs input, harmonic current in quadrature direction is obtained

With 0 as given value and harmonic current

Taking the difference and the result as the input of a seventh PI controller (64), the seventh PI controller (64) obtains the quadrature axis compensation voltage u_BqhWill control the voltage u in the quadrature axis direction_BqAnd quadrature axis compensation voltage u_BqhAdding to obtain quadrature axis command voltage

2. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 1, characterized in that: the first neural network band-pass filter (51) in the x direction comprises a first weight value adjusting module (55), and the actual displacement x and the vibration displacement

Difference is made to obtain error e_xError e_xMechanical angle theta to rotor_mThe sine and cosine values are used as the input of a first weight value adjusting module (55) to obtain the updated weight value omega in the x direction_{x_1}And ω_{x_2}(ii) a The second neural network band-pass filter (53) comprises a second weight adjusting module (56) which adjusts the actual displacement y and the vibration displacement

Differencing to obtain an error signal e_yWill error signal e_yMechanical angle theta to rotor_mThe sine and cosine values are used as the input of a second weight value adjusting module (56) to obtain the updated weight value omega in the y direction_{y_1}And ω_{y_2}。

3. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 2, characterized in that: vibration displacement at time k

Wherein, ω is_{x_1}(k+1)＝ω_{x_1}(k)+2μ₁e_xcosθ_m，ω_{x_2}(k+1)＝ω_{x_2}(k)+2μ₁e_xsinθ_m，ω_{x_1}，ω_{x_2}The weight value updated in the x direction; mu.s₁Is the step size factor.

4. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 1, characterized in that: the third neural network band-pass filter (61) comprises a third weight adjusting module (65) and a current i in the direction of the straight axis_BdAnd the harmonic current signal output by the third neural network band-pass filter (61)

Differencing to obtain a current error e_BdError of current e_BdAnd sin6 theta_e、cos6θ_eAs an input of a third weight adjustment module (65), obtaining an updated weight ω in the direction of the straight axis_{d6_1}And ω_{d6_2}The third neural network band-pass filter (61) outputs harmonic current

The fourth neural network band-pass filter (63) comprises a fourth weight value adjusting module (66) and a current i in the quadrature axis direction_BqAnd the harmonic current signal output by the fourth neural network band-pass filter (63)

Differencing to obtain a current error e_BqSignal, current error e_BqAnd sin6 theta_e、cos6θ_eAs the input of a fourth weight value adjusting module (66), the updated weight value omega in the direction of the straight axis is obtained_{q6_1}And ω_{q6_2}Output harmonic current

5. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 4, which is characterized in that: harmonic current at time k

Wherein, ω is_{d6_1}(k+1)＝ω_{d6_1}(k)+2μ₂e_Bdcos6θ_e，ω_{d6_2}(k+1)＝ω_{d6_2}(k)+2μ₂e_Bdsin6θ_e，ω_{d6_1}，ω_{d6_2}For the weight of the 6 th harmonic updated in the direction of the direct axis, mu₂Is the step size factor.

6. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 1, characterized in that: actual displacement x and given value x in x direction^*Obtaining a displacement error by difference making, inputting the error into a first PID controller (11), and obtaining a given value F of the force in the x direction of the suspension winding after being adjusted by the first PID controller (11)_x(ii) a The actual displacement y in the y direction is compared with a given value y^*The difference is made to obtain a displacement error, the displacement error is input into a second PID controller (12), and the given value F of the force in the y direction of the suspension winding is obtained after the adjustment of the second PID controller (12)_y。

7. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 1, characterized in that: collecting two-phase suspension winding current of a bearingless permanent magnet synchronous motor and collecting current i_1AAnd i_1CInput to a fourth coordinate transformation module (91), and the actual current i of the AC-DC axis of the suspension winding is obtained through the fourth coordinate transformation module (91)_BqAnd i_Bd(ii) a The given value of the quadrature-direct axis current

And

and the actual current i_BqAnd i_BdThe difference is respectively made and the result is inputted into the corresponding fourth PI controller (14) and the fifth PI controller (15), the fourth PI controller (14) outputs the control voltage u in the direction of the straight shaft_BdA fifth PI controller (15) outputs the control voltage u in the quadrature axis direction_Bq。

8. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 1, characterized in that: the said direct and alternating axis command voltage

And

the third coordinate transformation module (16) is used as the input of the third coordinate transformation module (16), and the third coordinate transformation module (16) outputs the suspension winding voltage u under the static coordinate system_BαAnd u_BβVoltage u of the levitation winding_BαAnd u_BβThe second SVPWM inverter (90) obtains a three-phase input voltage u of the bearingless permanent magnet synchronous motor as an input of the second SVPWM inverter (90)_1A、u_1B、u_1C。

9. According to the claimsSolving 1 the vibration compensation controller of the bearingless permanent magnet synchronous motor neural network band-pass filter is characterized in that: an encoder is adopted to collect pulse signals of the bearingless permanent magnet synchronous motor, and the mechanical angle of the rotor at the moment k is obtained through an angle calculation module (17)

Delta P is the accumulated result of the pulses output by the encoder, L_eIs the number of lines of the encoder.

10. The bearingless permanent magnet synchronous motor neural network band-pass filter vibration compensation controller of claim 9, wherein: rotor electrical angle theta at time k_e(k)＝P_Mθ_m(k)，θ_m(k) Mechanical angle of rotor at time k, P_MIs the torque winding pole pair number.